HackerPoet

This project is currently idle right now. I may come back to it when I get a chance to extend it to VR, though honestly I'll first try again to...

This network can only be trained on a GPU because it uses CuDNNGRU, the error sounds like you're trying to run it on a CPU. You may be able to...

Installing CUDA and cuDNN is best because it will train significantly faster. But if you can't get it working, just replace the instance of 'CuDNNGRU' with 'GRU'.

It may be possible, but it uses a lot of Windows APIs for other things like mouse/keyboard input, windowing, and sound/music (future). If you're interested in porting, I can provide...

I tried that originally, but it makes the songs too busy. There may be some middle-ground, but I'm not actively working on this project at the moment.

How stable would you say this PR is right now? I haven't been able to test it myself, but if you were able to parse the data, train a network,...

That's typical when all sliders are zeroed at the center. Does it still happen when you generate random songs with 'R' or adjust the sliders?

Assuming you had `WRITE_HISTORY` on, did any of the earlier epochs produce music that had notes outside the first beat?

That's strange. Any other info would be helpful. What did your loss look like? Mine was around 0.003 and converged quickly. If you're getting a similar loss, then it may...

Okay, that does looks like over-fitting to the smaller dataset. Try increasing the dropout rate `DO_RATE` to 0.25 or 0.5 to see if that helps. Also, just as a sanity...